4 Changes in version 1.75b:
6 * Bison now parses C99 lexical constructs like digraphs, UCNs, and
7 backslash-newline within C escape sequences, as POSIX now requires.
9 Changes in version 1.75a, 2002-10-24:
11 * Bison can no longer be built by a K&R C compiler; it requires C89 or
12 later to be built. This change originally took place a few versions
13 ago, but nobody noticed until we recently asked someone to try
14 building Bison with a K&R C compiler.
17 This new directive is preferred over YYERROR_VERBOSE.
19 * %lex-param, %parse-param
20 These new directives are preferred over PARSE_PARAM and LEX_PARAM.
21 In addition, they provide a means for yyerror to remain pure, and
22 to access to the current location.
24 Changes in version 1.75, 2002-10-14:
26 * Bison should now work on 64-bit hosts.
28 * Indonesian translation thanks to Tedi Heriyanto.
31 Fix spurious parse errors.
34 Some people redefine yyerror to steal yyparse' private variables.
35 Reenable this trick until an official feature replaces it.
38 In agreement with POSIX and with other Yaccs, leaving a default
39 action is valid when $$ is untyped, and $1 typed:
43 but the converse remains an error:
47 * Values of mid-rule actions
50 foo: { ... } { $$ = $1; } ...
52 was incorrectly rejected: $1 is defined in the second mid-rule
53 action, and is equal to the $$ of the first mid-rule action.
55 Changes in version 1.50, 2002-10-04:
60 causes Bison to produce a Generalized LR (GLR) parser, capable of handling
61 almost any context-free grammar, ambiguous or not. The new declarations
62 %dprec and %merge on grammar rules allow parse-time resolution of
63 ambiguities. Contributed by Paul Hilfinger.
65 Unfortunately Bison 1.50 does not work properly on 64-bit hosts
66 like the Alpha, so please stick to 32-bit hosts for now.
69 When not in Yacc compatibility mode, when the output file was not
70 specified, running `bison foo/bar.y' created `foo/bar.c'. It
74 The undefined token was systematically mapped to 2 which prevented
75 the use of 2 by the user. This is no longer the case.
77 * Unknown token numbers
78 If yylex returned an out of range value, yyparse could die. This is
82 According to POSIX, the error token must be 256.
83 Bison extends this requirement by making it a preference: *if* the
84 user specified that one of her tokens is numbered 256, then error
85 will be mapped onto another number.
87 * Verbose error messages
88 They no longer report `..., expecting error or...' for states where
89 error recovery is possible.
92 Defaults to `$end' instead of `$'.
94 * Error recovery now conforms to documentation and to POSIX
95 When a Bison-generated parser encounters a syntax error, it now pops
96 the stack until it finds a state that allows shifting the error
97 token. Formerly, it popped the stack until it found a state that
98 allowed some non-error action other than a default reduction on the
99 error token. The new behavior has long been the documented behavior,
100 and has long been required by POSIX. For more details, please see
101 <http://mail.gnu.org/pipermail/bug-bison/2002-May/001452.html>.
104 Popped tokens and nonterminals are now reported.
107 Larger grammars are now supported (larger token numbers, larger grammar
108 size (= sum of the LHS and RHS lengths), larger LALR tables).
109 Formerly, many of these numbers ran afoul of 16-bit limits;
110 now these limits are 32 bits on most hosts.
112 * Explicit initial rule
113 Bison used to play hacks with the initial rule, which the user does
114 not write. It is now explicit, and visible in the reports and
118 Before, Bison reported the useless rules, but, although not used,
119 included them in the parsers. They are now actually removed.
121 * Useless rules, useless nonterminals
122 They are now reported, as a warning, with their locations.
124 * Rules never reduced
125 Rules that can never be reduced because of conflicts are now
128 * Incorrect `Token not used'
131 %token useless useful
133 exp: '0' %prec useful;
135 where a token was used to set the precedence of the last rule,
136 bison reported both `useful' and `useless' as useless tokens.
138 * Revert the C++ namespace changes introduced in 1.31
139 as they caused too many portability hassles.
142 By an accident of design, the default computation of @$ was
143 performed after another default computation was performed: @$ = @1.
144 The latter is now removed: YYLLOC_DEFAULT is fully responsible of
145 the computation of @$.
148 The token end of file may be specified by the user, in which case,
149 the user symbol is used in the reports, the graphs, and the verbose
150 error messages instead of `$end', which remains being the default.
154 %token YYEOF 0 "end of file"
157 This old option, which has been broken for ages, is removed.
160 Brazilian Portuguese, thanks to Alexandre Folle de Menezes.
161 Croatian, thanks to Denis Lackovic.
163 * Incorrect token definitions
164 When given `%token 'a' "A"', Bison used to output `#define 'a' 65'.
166 * Token definitions as enums
167 Tokens are output both as the traditional #define's, and, provided
168 the compiler supports ANSI C or is a C++ compiler, as enums.
169 This lets debuggers display names instead of integers.
172 In addition to --verbose, bison supports --report=THINGS, which
173 produces additional information:
175 complete the core item sets with their closure
177 explicitly associate lookaheads to items
179 describe shift/reduce conflicts solving.
180 Bison used to systematically output this information on top of
181 the report. Solved conflicts are now attached to their states.
184 Previous versions don't complain when there is a type clash on
185 the default action if the rule has a mid-rule action, such as in:
193 * GNU M4 is now required when using Bison.
195 Changes in version 1.35, 2002-03-25:
198 Some projects use Bison's C parser with C++ compilers, and define
199 YYSTYPE as a class. The recent adjustment of C parsers for data
200 alignment and 64 bit architectures made this impossible.
202 Because for the time being no real solution for C++ parser
203 generation exists, kludges were implemented in the parser to
204 maintain this use. In the future, when Bison has C++ parsers, this
205 kludge will be disabled.
207 This kludge also addresses some C++ problems when the stack was
210 Changes in version 1.34, 2002-03-12:
212 * File name clashes are detected
213 $ bison foo.y -d -o foo.x
214 fatal error: header and parser would both be named `foo.x'
216 * A missing `;' at the end of a rule triggers a warning
217 In accordance with POSIX, and in agreement with other
218 Yacc implementations, Bison will mandate this semicolon in the near
219 future. This eases the implementation of a Bison parser of Bison
220 grammars by making this grammar LALR(1) instead of LR(2). To
221 facilitate the transition, this release introduces a warning.
223 * Revert the C++ namespace changes introduced in 1.31, as they caused too
224 many portability hassles.
226 * DJGPP support added.
228 * Fix test suite portability problems.
230 Changes in version 1.33, 2002-02-07:
233 Groff could not be compiled for the definition of size_t was lacking
234 under some conditions.
239 Changes in version 1.32, 2002-01-23:
241 * Fix Yacc output file names
245 * Italian, Dutch translations
247 Changes in version 1.31, 2002-01-14:
251 * GNU Gettext and %expect
252 GNU Gettext asserts 10 s/r conflicts, but there are 7. Now that
253 Bison dies on incorrect %expectations, we fear there will be
254 too many bug reports for Gettext, so _for the time being_, %expect
255 does not trigger an error when the input file is named `plural.y'.
257 * Use of alloca in parsers
258 If YYSTACK_USE_ALLOCA is defined to 0, then the parsers will use
259 malloc exclusively. Since 1.29, but was not NEWS'ed.
261 alloca is used only when compiled with GCC, to avoid portability
264 * When the generated parser lacks debugging code, YYDEBUG is now 0
265 (as POSIX requires) instead of being undefined.
268 Bison has always permitted actions such as { $$ = $1 }: it adds the
269 ending semicolon. Now if in Yacc compatibility mode, the semicolon
270 is no longer output: one has to write { $$ = $1; }.
272 * Better C++ compliance
273 The output parsers try to respect C++ namespaces.
274 [This turned out to be a failed experiment, and it was reverted later.]
277 Fixed bugs when reporting useless nonterminals.
280 The parsers work properly on 64 bit hosts.
283 Some calls to strerror resulted in scrambled or missing error messages.
286 When the number of shift/reduce conflicts is correct, don't issue
289 * The verbose report includes the rule line numbers.
291 * Rule line numbers are fixed in traces.
293 * Swedish translation
296 Verbose parse error messages from the parsers are better looking.
297 Before: parse error: unexpected `'/'', expecting `"number"' or `'-'' or `'(''
298 Now: parse error: unexpected '/', expecting "number" or '-' or '('
300 * Fixed parser memory leaks.
301 When the generated parser was using malloc to extend its stacks, the
302 previous allocations were not freed.
304 * Fixed verbose output file.
305 Some newlines were missing.
306 Some conflicts in state descriptions were missing.
308 * Fixed conflict report.
309 Option -v was needed to get the result.
313 Mismatches are errors, not warnings.
315 * Fixed incorrect processing of some invalid input.
317 * Fixed CPP guards: 9foo.h uses BISON_9FOO_H instead of 9FOO_H.
319 * Fixed some typos in the documentation.
321 * %token MY_EOF 0 is supported.
322 Before, MY_EOF was silently renumbered as 257.
324 * doc/refcard.tex is updated.
326 * %output, %file-prefix, %name-prefix.
330 New, aliasing `--output-file'.
332 Changes in version 1.30, 2001-10-26:
334 * `--defines' and `--graph' have now an optionnal argument which is the
335 output file name. `-d' and `-g' do not change, they do not take any
338 * `%source_extension' and `%header_extension' are removed, failed
343 Changes in version 1.29, 2001-09-07:
345 * The output file does not define const, as this caused problems when used
346 with common autoconfiguration schemes. If you still use ancient compilers
347 that lack const, compile with the equivalent of the C compiler option
348 `-Dconst='. autoconf's AC_C_CONST macro provides one way to do this.
350 * Added `-g' and `--graph'.
352 * The Bison manual is now distributed under the terms of the GNU FDL.
354 * The input and the output files has automatically a similar extension.
356 * Russian translation added.
358 * NLS support updated; should hopefully be less troublesome.
360 * Added the old Bison reference card.
362 * Added `--locations' and `%locations'.
364 * Added `-S' and `--skeleton'.
366 * `%raw', `-r', `--raw' is disabled.
368 * Special characters are escaped when output. This solves the problems
369 of the #line lines with path names including backslashes.
372 `%yacc', `%fixed_output_files', `%defines', `%no_parser', `%verbose',
373 `%debug', `%source_extension' and `%header_extension'.
376 Automatic location tracking.
378 Changes in version 1.28, 1999-07-06:
380 * Should compile better now with K&R compilers.
384 * Fixed a problem with escaping the double quote character.
386 * There is now a FAQ.
388 Changes in version 1.27:
390 * The make rule which prevented bison.simple from being created on
391 some systems has been fixed.
393 Changes in version 1.26:
395 * Bison now uses automake.
397 * New mailing lists: <bug-bison@gnu.org> and <help-bison@gnu.org>.
399 * Token numbers now start at 257 as previously documented, not 258.
401 * Bison honors the TMPDIR environment variable.
403 * A couple of buffer overruns have been fixed.
405 * Problems when closing files should now be reported.
407 * Generated parsers should now work even on operating systems which do
408 not provide alloca().
410 Changes in version 1.25, 1995-10-16:
412 * Errors in the input grammar are not fatal; Bison keeps reading
413 the grammar file, and reports all the errors found in it.
415 * Tokens can now be specified as multiple-character strings: for
416 example, you could use "<=" for a token which looks like <=, instead
417 of chosing a name like LESSEQ.
419 * The %token_table declaration says to write a table of tokens (names
420 and numbers) into the parser file. The yylex function can use this
421 table to recognize multiple-character string tokens, or for other
424 * The %no_lines declaration says not to generate any #line preprocessor
425 directives in the parser file.
427 * The %raw declaration says to use internal Bison token numbers, not
428 Yacc-compatible token numbers, when token names are defined as macros.
430 * The --no-parser option produces the parser tables without including
431 the parser engine; a project can now use its own parser engine.
432 The actions go into a separate file called NAME.act, in the form of
433 a switch statement body.
435 Changes in version 1.23:
437 The user can define YYPARSE_PARAM as the name of an argument to be
438 passed into yyparse. The argument should have type void *. It should
439 actually point to an object. Grammar actions can access the variable
440 by casting it to the proper pointer type.
442 Line numbers in output file corrected.
444 Changes in version 1.22:
448 Changes in version 1.20:
450 Output file does not redefine const for C++.
458 Copyright (C) 2001, 2002 Free Software Foundation, Inc.
460 This file is part of GNU Autoconf.
462 GNU Autoconf is free software; you can redistribute it and/or modify
463 it under the terms of the GNU General Public License as published by
464 the Free Software Foundation; either version 2, or (at your option)
467 GNU Autoconf is distributed in the hope that it will be useful,
468 but WITHOUT ANY WARRANTY; without even the implied warranty of
469 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
470 GNU General Public License for more details.
472 You should have received a copy of the GNU General Public License
473 along with autoconf; see the file COPYING. If not, write to
474 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
475 Boston, MA 02111-1307, USA.